a2-3
April 25th, 2016, 12:07 AM
Hello.
I used the weekend to purge my Xubuntu 14.04 installations on three machines and to install a fresh 16.04 on them. To be honest: I'm deeply disappointed. I already heard that development for the desktop world stalled in favour to other battlefield. Well, okay, this is a different story... I mean, a stable OS is not a bad thing. When things are mature and only change slightly, I'm really fine with that. But what xenial brought, compared to trusty was nothing more than a pile of regressions.
In trusty, I've seen lots of minor flaws. Not a single one is fixed. Not one. I mean, it begins with the installer with the not-that-perfect idea to let the user make partitioning before asking for the keyboard layout. Not a breaking issue, I admit. But if Redmond had came up with such an idea, we all would have laughed loudly about it. It was not fixed in the last two years. The UI style differs radically among Qt4/Qt5/Gtk2/Gtk3. There is no single convegent settings UI, which covers everything. There is no solid support for non-96dpi-screens. All workstation infrastructure (sound, networking, bluetooth, login-sessions, filesystem-mounts, ...) is somewhat buggy and serves you with random error phenomena instead of real operation if you click on the wrong stuff, or too fast, or in a bad order, or in the wrong situation. For some very common configuration aspects, no mature UI exist at all (e.g. filetype associations) in a default installation. There are tons of minor issues in trusty.
But it's not only that devs were not able to fix any of that bugs I saw in the last two years. They actually managed to bring in annoying new ones. My screen flickers weirdly at boot time. After password input, the 'Please enter password' reappears for half a second. My mouse cursor disappears when I lock the screen. The infrastructure is not able to install .deb packages graphically (!!!!) but just via good old dpkg. Some things are noticeable slower than before (how was that realized without really bringing anything new??). The xfce terminal window becomes blank black when I move it to another screen. And so forth and so forth... And this is just what I have seen after an hour or two.
I am very aware of the fact that many issues do not come from the Ubuntu team but from the various projects (Xfce, freedesktop, firefox, systemd, libreoffice, drivers, whatever). So this is not only about Ubuntu but maybe also about the FOSS ecosystem in general (if Ubuntu really sees itself as an actual part of it - I have some doubts after stuff like Mir, Unity, Snappy, ...). I love it that people voluntarily write software. I know that I don't have any legal claim for them to do anything. But why the hell is the whole community nuts about implementing obviously not-that-perfect stuff like integrations to non-free cloud services and not-really-working mobile convergence, while most of the infrastructure have (thereby increasing) flaws in its basic functionality?
Is this a problem which does not really exist and I just imagine? Or is this considered as an actual issue by other people as well? Are there any idea about countermeasures? I mean, of course it is mentioned to be an option to participate in all that projects by writing to mailing lists, filing bugs, writing patches, ... - and I admit that this could really work. But that would mean to spend a whole lifetime by doing other peoples' QA. And this is not as easy as it sounds. Ubuntu versions are always that old that bug reports to upstream get ignored. Fixing a bug in the source is merely a theoretical option, since many bugs are results of wrong decisions in the lowest-level infrastructure of a software (and lots of dirty not-really-working hacks on top of that). This is neither easy to fix nor easy to actually bring to upstream (I admit, if people would send me patches about my programs, which bring deep infrastructure changes, I would also be somewhat reserved).
This post should not only be some kind of bitching. Sorry, if it sounds this way. What are your strategies for participating somehow to FOSS software quality? I'm a bit helpless at the moment. I would do something, if I had the impression that it actually moves things forward. But all ways I know about are only some kind of occupational therapy, or they are not really seizable for non-project-members. What is a good way to patch Ubuntu's xfce (just for instance) without turning the entire system inside out? What is a good way to tell Gnome people that things like 'Client Side Decorations' are a really really bad idea? I mean, it break that much things (most of them admittedly not relevant for day-to-day usage, but also some relevant stuff about consistency) just for 'being modern'? And this is just one example which comes to my mind, while I see a very general trend towards those sad ideas.
For a minimal answer, please just tell me: Do you see this as a problem as well? Is there a general trend towards various ill directions while meaningful development largely stalled? Or have I developed an ill perspective of observation?
Greetings
Josef
PS: Dear admin: Hopefully it is now okay enough for you to not censor again. The original post contained some corollary adjectives which are not 'flowers and bees'-like and came from my frustration. I replaced them with flowers and bees. Due to broken forum software (only works with Javascript enabled) it also did not contain linebreaks. Immediate post removal for non-unlawful content is not entirely what I would expect from such a forum. Maybe, when you eventually remove it again, you could point me to the actual issue.
I used the weekend to purge my Xubuntu 14.04 installations on three machines and to install a fresh 16.04 on them. To be honest: I'm deeply disappointed. I already heard that development for the desktop world stalled in favour to other battlefield. Well, okay, this is a different story... I mean, a stable OS is not a bad thing. When things are mature and only change slightly, I'm really fine with that. But what xenial brought, compared to trusty was nothing more than a pile of regressions.
In trusty, I've seen lots of minor flaws. Not a single one is fixed. Not one. I mean, it begins with the installer with the not-that-perfect idea to let the user make partitioning before asking for the keyboard layout. Not a breaking issue, I admit. But if Redmond had came up with such an idea, we all would have laughed loudly about it. It was not fixed in the last two years. The UI style differs radically among Qt4/Qt5/Gtk2/Gtk3. There is no single convegent settings UI, which covers everything. There is no solid support for non-96dpi-screens. All workstation infrastructure (sound, networking, bluetooth, login-sessions, filesystem-mounts, ...) is somewhat buggy and serves you with random error phenomena instead of real operation if you click on the wrong stuff, or too fast, or in a bad order, or in the wrong situation. For some very common configuration aspects, no mature UI exist at all (e.g. filetype associations) in a default installation. There are tons of minor issues in trusty.
But it's not only that devs were not able to fix any of that bugs I saw in the last two years. They actually managed to bring in annoying new ones. My screen flickers weirdly at boot time. After password input, the 'Please enter password' reappears for half a second. My mouse cursor disappears when I lock the screen. The infrastructure is not able to install .deb packages graphically (!!!!) but just via good old dpkg. Some things are noticeable slower than before (how was that realized without really bringing anything new??). The xfce terminal window becomes blank black when I move it to another screen. And so forth and so forth... And this is just what I have seen after an hour or two.
I am very aware of the fact that many issues do not come from the Ubuntu team but from the various projects (Xfce, freedesktop, firefox, systemd, libreoffice, drivers, whatever). So this is not only about Ubuntu but maybe also about the FOSS ecosystem in general (if Ubuntu really sees itself as an actual part of it - I have some doubts after stuff like Mir, Unity, Snappy, ...). I love it that people voluntarily write software. I know that I don't have any legal claim for them to do anything. But why the hell is the whole community nuts about implementing obviously not-that-perfect stuff like integrations to non-free cloud services and not-really-working mobile convergence, while most of the infrastructure have (thereby increasing) flaws in its basic functionality?
Is this a problem which does not really exist and I just imagine? Or is this considered as an actual issue by other people as well? Are there any idea about countermeasures? I mean, of course it is mentioned to be an option to participate in all that projects by writing to mailing lists, filing bugs, writing patches, ... - and I admit that this could really work. But that would mean to spend a whole lifetime by doing other peoples' QA. And this is not as easy as it sounds. Ubuntu versions are always that old that bug reports to upstream get ignored. Fixing a bug in the source is merely a theoretical option, since many bugs are results of wrong decisions in the lowest-level infrastructure of a software (and lots of dirty not-really-working hacks on top of that). This is neither easy to fix nor easy to actually bring to upstream (I admit, if people would send me patches about my programs, which bring deep infrastructure changes, I would also be somewhat reserved).
This post should not only be some kind of bitching. Sorry, if it sounds this way. What are your strategies for participating somehow to FOSS software quality? I'm a bit helpless at the moment. I would do something, if I had the impression that it actually moves things forward. But all ways I know about are only some kind of occupational therapy, or they are not really seizable for non-project-members. What is a good way to patch Ubuntu's xfce (just for instance) without turning the entire system inside out? What is a good way to tell Gnome people that things like 'Client Side Decorations' are a really really bad idea? I mean, it break that much things (most of them admittedly not relevant for day-to-day usage, but also some relevant stuff about consistency) just for 'being modern'? And this is just one example which comes to my mind, while I see a very general trend towards those sad ideas.
For a minimal answer, please just tell me: Do you see this as a problem as well? Is there a general trend towards various ill directions while meaningful development largely stalled? Or have I developed an ill perspective of observation?
Greetings
Josef
PS: Dear admin: Hopefully it is now okay enough for you to not censor again. The original post contained some corollary adjectives which are not 'flowers and bees'-like and came from my frustration. I replaced them with flowers and bees. Due to broken forum software (only works with Javascript enabled) it also did not contain linebreaks. Immediate post removal for non-unlawful content is not entirely what I would expect from such a forum. Maybe, when you eventually remove it again, you could point me to the actual issue.